Search results

The progress of parallel computing in Information Retrieval (IR) is reviewed. In particular we stress the importance of the motivation in using parallel computing for text retrieval. We analyse parallel IR systems using a classification defined by Rasmussen and describe some parallel IR systems. We give a description of the retrieval models used in parallel information processing. We describe areas of research which we believe are needed.

The proposed algorithm successfully optimizes complex error functions, which are difficult to differentiate, ill conditioned or discontinuous. It is a benchmark to identify initial solutions in artificial neural network (ANN) training.

A multi‐directional ANN training algorithm that needs no derivative information is introduced as constrained one‐dimensional problem. A directional search vector examines the ANN error function in weight parameter space. The search vector moves in all possible directions to find minimum function value. The network weights are increased or decreased depending on the shape of the error function hyper surface such that the search vector finds descent directions. The minimum function value is thus determined. To accelerate the convergence of the algorithm a momentum search is designed. It avoids overshooting the local minimum.

The training algorithm is insensitive to the initial starting weights in comparison with the gradient‐based methods. Therefore, it can locate a relative local minimum from anywhere of the error surface. It is an important property of this training method. The algorithm is suitable for error functions that are discontinuous, ill conditioned or the derivative of the error function is not readily available. It improves over the standard back propagation method in convergence and avoids premature termination near pseudo local minimum.

Classifications problems are efficiently classified when using this method but the complex time series in some instances slows convergence due to complexity of the error surface. Different ANN network structure can further be investigated to find the performance of the algorithm.

The search scheme moves along the valleys and ridges of the error function to trace minimum neighborhood. The algorithm only evaluates the error function. As soon as the algorithm detects flat surface of the error function, care is taken to avoid slow convergence.

The algorithm is efficient due to incorporation of three important methodologies. The first mechanism is the momentum search. The second methodology is the implementation of directional search vector in coordinate directions. The third procedure is the one‐dimensional search in constrained region to identify the self‐adaptive learning rates, to improve convergence.

This paper sets out to evaluate the effect of search method employed (broadly defined here as formal or informal) on the probability that an individual will move out of unemployment into either employment or inactivity. We take advantage of the longitudinal element contained within the Labour Force Survey and link 794 individual records through five quarters. The sample contains all those who are unemployed in Spring 1996 and records their labour market status for the four quarters ending in Spring 1997. Given that this element of the LFS is only just beginning to be used, we attempt to evaluate its usefulness for the estimation of a discrete time‐competing risks model. We do find an effect of search method employed in the final specification and the sign on the coefficient rejects our belief that those using informal search methods may do so in order to signal their higher levels of productivity. Instead we are led to conclude that those using informal search methods may do so because of a lack of skills needed to use more formal channels.

The Labour Force Survey is a rich data source for investigation of the factors influencing unemployment. The initial results are based on a linear probability model. They highlight some of the influences on job search, the choice of methods, the intensity and the duration of search. The probability of search is positively related to claimancy status and is higher for males and for those in younger‐age and shorter‐duration unemployment categories. Family responsibilities had quite different effects on male and female job search probabilities, except in the case of single parents. The intensity of search was noticeably higher in males, for claimants, for younger‐age groups and shorter‐duration categories. Reliance on job centres fell significantly with age, tended to be higher amongst the less well qualified and lower social classes. With the exception of Northern Ireland, individuals in higher unemployment areas made greater use of job centres. All of the results were duration‐related: the probability and intensity of job search fell with the duration of unemployment; the reliance on job centres increased with duration.

The purpose of this paper is to examine the effectiveness of job search and human capital of the unemployed in the Russian Federation for obtaining a job the following year.

Cross‐sectional data on human capital and job‐search strategies from different years of the Russian Longitudinal Monitoring Survey are used in different probit specifications to present and analyse empirical results.

Searching through friends and directly are the most common search methods, implying a large trust in networks. Moreover, people tend to do the right thing when trusting their networks; in 1994, searching via friends and directly were the only methods with a positive impact. People used fewer methods in 1994 and the impact of job searching is also higher in the latter two samples. This indicates a society gradually adapting to unemployment. As regards human capital, it is worth noticing that the results in 2004 are in accordance with several predictions of human capital theory, whereas the impact of medium education and work experience in the samples based on earlier years is weaker. This suggests that more people have had time to upgrade their skills, to obtain relevant work experience and/or to obtain a post transition secondary education in 2003.

This paper identifies the importance of both different job‐search strategies and human capital when searching for a job in the Russian Federation. The result presented may be of interest to both policy‐makers and scientists.

The paper aims to improve convergence characteristics of the Newton–Raphson (NR) method applied to time-periodic finite element method using various line searches, as time-periodic finite element method causes deterioration of convergence characteristic of nonlinear analysis based on NR method. The study also aims to accelerate and improve accuracy of electromagnetic field analysis for improvement of the performance of electrical machine.

The paper proposes new type line searches that set approximate step size for NR method. The line search evaluated step size using higher-order interpolation of functional derivative. In addition, two criteria for applying these line search were proposed. First method set one scalar value for every NR iteration that is named constant step size. Second method define different step size in each time step of time-periodic finite element method to update solution vector that is named different step size.

The paper provides efficient line searches to improve convergence characteristics for NR method. Nonlinear magnetic field analysis of two transformer models is demonstrated. The proposed methods achieve the following results: higher-order functional NR is efficient in improving convergence characteristics, and the proposed methods succeeded about twice faster in both models.

The paper fulfills improvement of convergence characteristics of the NR method applied to time-periodic finite element method using proposed line searches and accelerate electromagnetic field analysis.

The purpose of this paper is to present a solution space searching method to study the initial design of interplanetary low thrust gravity assist trajectory.

For reducing the complexity and nonlinearity of the initial design problem, a sixth degree inverse polynomial shape based approach is brought. Then some improvements are provided for solving the parameters in the shape function and a quasi‐lambert solver is brought through the shape based method, the thrust profile can be generated under the given time of flight, boundary states including positions and velocities for low thrust phase. Combining gravity assist model, the problem is summarized and an improved pruning technique is used for searching the feasible solution space for low thrust gravity assist trajectory.

Using the solution space searching method, the feasible solution region would be generated under the given mission condition. The treatment about gravity assist demonstrates more accurate than previous method. Also another advantage is that the searching method can be used to design different types of mission trajectory, including flyby and rendezvous trajectories.

The method can be used as an efficient approach to search the feasible region for the complex low thrust gravity assist trajectory, and it can provide appropriate initial guesses for the low thrust gravity assist trajectory in mission design phase.

Feasible solution space would be obtained through the searching method. The quasi‐Lambert solver in the paper is found under the shape‐based method and relative improvement, and it shows its availability during the searching process. Through mission trajectory design, the effectiveness of the method is shown.

This paper aims to propose an environmentally adaptive trajectory planning system considering the dynamic characteristics of unmanned aerial vehicles (UAVs) and the distance between obstacles and the UAV. The system generates a smooth and safe flight trajectory online.

First, the hybrid A* search method considering the dynamic characteristics of the quadrotor is used to find the collision-free initial trajectory. Then, environmentally adaptive velocity cost is designed for environment-adaptive trajectory optimization using environmental gradient data. The proposed method adaptively adjusts the autonomous flight speed of the UAV. Finally, the initial trajectory is applied to the multi-layered optimization framework to make it smooth and dynamically viable.

The feasibility of the designed system is validated by online flight experiments, which are in unknown, complex situations.

The proposed trajectory planning system is integrated into a vision-based quadrotor platform. It is easily implementable onboard and computationally efficient.

A hybrid A* path searching method is proposed to generate feasible motion primitives by dispersing the input space uniformly. The proposed method considers the control input of the UAV and the search time as the heuristic cost. Therefore, the proposed method can provide an initial path with the minimum flying time and energy loss that benefits trajectory optimization. The environmentally adaptive velocity cost is proposed to adaptively adjust the flight speed of the UAV using the distance between obstacles and the UAV. Furthermore, a multi-layered environmentally adaptive trajectory optimization framework is proposed to generate a smooth and safe trajectory.

This study aims to present a new web page recommendation system that can help users to reduce navigational time on the internet.

The proposed design is based on the primacy effect of browsing behavior, that users prefer top ranking items in search results. This approach is intuitive and requires no training data at all.

A user study showed that users are more satisfied with the proposed search methods than with general search engines using hot keywords. Moreover, two performance measures confirmed that the proposed search methods out‐perform other metasearch and search engines.

The research has limitations and future work is planned along several directions. First, the search methods implemented are primarily based on the keyword match between the contents of web pages and the user query items. Using the semantic web to recommend concepts and items relevant to the user query might be very helpful in finding the exact contents that users want, particularly when the users do not have enough knowledge about the domains in which they are searching. Second, offering a mechanism that groups search results to improve the way search results are segmented and displayed also assists users to locate the contents they need. Finally, more user feedback is needed to fine‐tune the search parameters including α and β to improve the performance.

The proposed model can be used to improve the search performance of any search engine.

First, compared with the democratic voting procedure used by metasearch engines, search engine vector voting (SVV) enables a specific combination of search parameters, denoted as α and β, to be applied to a voted search engine, so that users can either narrow or expand their search results to meet their search preferences. Second, unlike page quality analysis, the hyperlink prediction (HLP) determines qualified pages by simply measuring their user behavior function (UBF) values, and thus takes less computing power. Finally, the advantages of HLP over statistical analysis are that it does not need training data, and it can target both multi‐site and site‐specific analysis.

Many search engines in digital libraries are restricted to the terms presented in users’ queries. When users cannot represent their information needs in terms of keywords in a query, the search engine fails to provide appropriate results. In addition, most search engines do not have the ability to visualize search results for users to help them in their information journey. The purpose of this paper is to develop a new approach for search result visualization in digital libraries. The visualization approach enables subject-based visualization of search results and search queries.

To enable subject-based visualization of search results in digital libraries, new subject-based document retrieval is proposed in which each document is represented as a vector of subjects as well. Then, using a vector space model for information retrieval, along with the subject-based vector, related documents to the user’s query are retrieved, whilst each document is visualized through a ring chart, showing the inherent subjects within each document and the query.

The proposed subject-based retrieval and visualization approach is evaluated from various perspectives to amplify the impact of the visualization approach from users’ opinions. Users have evaluated the performance of the proposed subject-based retrieval and search result visualization, whilst 67% of users prefer subject-based document retrieval and 80% of them believe that the proposed visualization approach is practical.

This research has provided a subject-based representation scheme for search result visualization in a digital library. The implication of this research can be viewed from two perspectives. First, the subject-based retrieval approach provides an opportunity for the users to understand their information needs, beyond the explicit terms in the query, leading to results, which are semantically relevant to the query. Second, the simple subject-based visualization scheme, helps users to explore the results easily, whilst allowing them to build their knowledge experience.

A new vectorized subject-based representation of documents and queries is proposed. This representation determines the semantic and subject-based relationship between a given query and documents within a digital scientific library. In addition, it also provides a subject-based representation of the retrieved documents through which users can track the semantic relationship between the query and retrieve documents, visually.